Live Caption Pro
Menu

Screenshots

What the product looks like: the caption band over a call, the caption window with its toolbar and settings panels, every tab of the settings window, the Recordings window and the web administration of the services. Every caption says how the picture was made.

Click a picture for the large version.

The caption window

The caption window: the first picture is a drawn illustration (both the call window and the band are drawn); the others are the program's real caption window with its toolbar and settings panels, rendered by machine with sample text.

Illustration: an abstract call window with the caption band over it, showing a sentence and its translation.
How it sits over a call. Both the call window and the band are drawn; the band follows the look of the program's caption band. Illustration.
The caption window with two-language captions, and the four toolbar groups below: incoming, translation, outgoing, window.
The toolbar slides in under the band on hover; each coloured group switches one direction, and its gear opens that group's settings panel. Rendered by machine from the macOS client's interface, with sample text.
The Incoming audio panel: audio source and device, text size, recognition profile and mode, sentence boundary and more.
The incoming panel: where the audio comes from, which engine transcribes it, and how the caption is split. Rendered by machine from the macOS client's interface, with sample text.
The Translation and interpreting panel: language pair, text size, playback device, volumes, the voice and the interpreter.
The translation panel: the target language, the interpreter's voice and its pace. Rendered by machine from the macOS client's interface, with sample text.
The Outgoing panel: captions of your own speech, the microphone, the cable, the volumes sent into the call, the interpreter.
The outgoing panel: your own speech is captioned and translated too, and the interpreter's voice goes into the call. Rendered by machine from the macOS client's interface, with sample text.
The Caption window panel: background colour, opacity, blur, spacing, the translation's position and when captions fade.
The window panel: the band's background and how old captions fade; the other window settings are in the settings window. Rendered by machine from the macOS client's interface, with sample text.

The settings window

The settings window and the Recordings window: the program's real windows, rendered by machine from the macOS client's interface, with test data.

The Usage tab: the mode picker, the call app check and the audio path table with the devices.
The Usage tab: one choice sets the audio path for the whole session; the table shows where the program listens. Rendered by machine from the macOS client's interface, with sample text.
The Audio cable tab: what the virtual audio cable is, and a diagram of the two directions with BlackHole cables.
The Audio cable tab: the interpreter's voice reaches the call over a cable; the program hears the call's audio without one too, through system audio capture (loopback) – a cable is only needed if the calling app plays to it. On macOS the cable is BlackHole. Rendered by machine from the macOS client's interface, with sample text.
The Live Captions tab: the current state of macOS's own speech recognition on this machine, and the steps to turn it on.
The Live Captions tab: macOS's own recognition gives captions without a server or a graphics card – but not in Hungarian. Rendered by machine from the macOS client's interface, with sample text.
The Servers tab: the server profile on top, below it the caption server's address, key field and web admin interface.
The Servers tab: an address and a key per service, saved in profiles; a local service can be started from here. Rendered by machine from the macOS client's interface, with sample text.
The Servers tab with the Help panel open: what you set here, the table of servers and the profiles explained.
The Help panel: the top of every tab explains that tab's own topic, right above the fields. Rendered by machine from the macOS client's interface, with sample text.
The Audio source tab: the source of the captions, the sound to caption, silent playback and the device picker.
The Audio source tab: where the captions and the audio come from – the speaker output, the microphone or the audio cable. Rendered by machine from the macOS client's interface, with sample text.
The Recordings tab: the recording checkbox, the days to keep and the recordings folder.
The Recordings tab: this is where recording is switched on, and where you set how long recordings are kept. Rendered by machine from the macOS client's interface, with sample text.
The Session tab: the recognition model, the segmentation mode, the three timing values and passing the previous sentence.
The Session tab: the recognition model, and how speech is split into sentences. Rendered by machine from the macOS client's interface, with sample text.
The Appearance tab: the interface and log language, then the caption window's opacity, blur, background and band spacing.
The Appearance tab: the interface language and the look of the caption band – colours, fonts, bands and transparency. Rendered by machine from the macOS client's interface, with sample text.
The Advanced tab: the speech detection threshold and the interval mode values, per direction.
The Advanced tab: fine-tuning speech detection and the three values of the interval mode, per direction. Rendered by machine from the macOS client's interface, with sample text.
The Translation tab: the translation switch, sentences translated at once, the two languages and spoken interpreting.
The Translation tab: the speech and target languages, and the interpreter that speaks the translation into your headphones. Rendered by machine from the macOS client's interface, with sample text.
The Shortcuts tab with the list of key combinations.
The Shortcuts tab: one field per action with the recorded combination. Rendered by machine from the macOS client's interface, with sample text.
The System tab: hiding the captions, the paths of the resolved environment, the open folder and copy buttons.
The System tab: how the program behaves, and which folder holds what – it can be copied to the clipboard for a fault report. Rendered by machine from the macOS client's interface, with sample text.
The Help tab with the description of the caption window, the toolbar, the menu and the shortcuts.
The Help tab: how to use the program, in one place inside the settings window. Rendered by machine from the macOS client's interface, with sample text.
The About tab: the version, the developer, the website, the four members of the product family and the licence.
The About tab: which release is running, and who makes the program. Rendered by machine from the macOS client's interface, with sample text.

Recordings

The settings window and the Recordings window: the program's real windows, rendered by machine from the macOS client's interface, with test data.

The Recordings window: the list of recordings on the left, the selected one's sentences with time, original and translation.
Recordings: a recorded session can be reviewed, searched and exported as subtitles afterwards. Rendered by machine from the macOS client's interface, with sample text.

Web administration

Web administration: a test environment with a stand-in engine, so the hardware and model data are not real.

The login page of a service's web administration, with the language picker.
The login page: the same for all four services. Test environment, stand-in engine.
The caption server's overview: engine state, loaded models, video memory, live sessions.
The caption server's overview page. Test environment, stand-in engine.
The Models page: the model catalogue with sizes, memory needs and fit.
The Models page: which models fit the card, and which are loaded. Test environment, stand-in engine.
The Decoding page: the Whisper decoding settings with their defaults.
The Decoding page: the decoding settings, each one optional. Test environment, stand-in engine.
The Connections page: the recent requests with their caller, status and duration.
The Connections page: who called, when, and how long it took. Test environment, stand-in engine.
The Callers page: the known clients with their permissions and state.
The Callers page: read-only; users and keys are managed in the account service. Test environment, stand-in engine.
The translator's overview: the engine, the active profile, the slots and the load curve.
The translator's overview page. Test environment, stand-in engine.
The Model profile page: model size, variant, cache type, slots, context and fit.
The Model profile page: a profile only starts if it provably fits the card. Test environment, stand-in engine.
The voice service's overview: the engine state, the model and the device.
The voice service's overview page. Test environment, stand-in engine.
The Voices page: the built-in voices and the stored voice profiles.
The Voices page: the built-in voices; cloned profiles come on top. Test environment, stand-in engine.
The Users page of the account service: users with roles and state.
The Users page: roles and passwords live here, for all services. Test environment, stand-in engine.
The Keys page: the API keys with their owner, expiry and state.
The Keys page: a key is shown once, at creation. Test environment, stand-in engine.
The Audit page: who changed what and when.
The Audit page: every change to users, roles and keys. Test environment, stand-in engine.

Light and dark theme

Web administration: a test environment with a stand-in engine, so the hardware and model data are not real.

The caption server's overview in the dark theme.
The dark theme: the same pages, chosen per browser. Test environment, stand-in engine.
Hiba történt az oldal futtatásakor. / An error occurred while running the page. Újratöltés / Reload 🗙